Voice AI cost estimator

Since 20 May 2026, HighLevel bills Voice AI in parts: a voice engine rate, a text-to-speech (voice) rate, LLM tokens, and phone minutes. This estimator adds them up and compares pay-per-use against the flat AI Employee plans.

Call volume
Voice and model
Input-token prices from HighLevel's table. Output tokens cost more, so treat this as a floor.
Our assumption, not a HighLevel figure. Long prompts and knowledge bases push it up. Check real usage after a week.

The rates used

ComponentRate
Voice engine$0.045/min (from 20 May 2026)
TTS: OpenAI or Cartesia$0.015/min
TTS: ElevenLabs V2.5$0.035/min
TTS: ElevenLabs V3$0.17/min
Speech-to-speech: Gemini live / OpenAI realtime$0.1 / $0.2 per min, flat
AI Employee Growth$50/mo per location, includes 100 Voice AI minutes; overage at pay-per-use
AI Employee Unlimited$97/mo per location, Voice AI unlimited subject to fair use
LC Phone, inbound US answered in app$0.01165/min, billed in whole minutes
LC Phone, outbound US$0.0166/min

HighLevel's formula: call minutes x (voice engine + TTS rate) + LLM token cost + phone charges. [AI Product Pricing (HighLevel help center), updated 2026-10-20; checked 2026-10-25] [LC Phone Pricing & Billing Guide, updated 2026-10-12; checked 2026-10-25]

What usually decides the cheapest option

Keep test costs down

When we built demo receptionist agents through the API, we set a maximum call duration (300 seconds) on each one so a stuck test call couldn't run up minutes. Web-call testing in the builder needs no phone number. See the Voice AI receptionist guide.

Start the 14-day HighLevel trial

Affiliate link: if you sign up through it we may earn a commission, at no extra cost to you. How we make money.

Facts last checked against official sources: